From what I have heard of plasma oxidation, the oxide of aluminum is
pretty self limiting unless you apply a voltage.  Even then you will only
get about 1.5-2 nm/volt.  You can also try anodization instead of plasma
oxidation with a similar thickness/voltage dependence.  Just make sure to
not use Sulfuric acid since that will give a porous oxide.

Glen

At 12:01 PM 6/19/2002, you wrote:
>Message: 4
>Date: Tue, 18 Jun 2002 00:09:21 -0400
>From: Nicholas Jankowski 
>To: [email protected]
>Subject: [mems-talk] Thermal growth rates of Aluminum Oxide
>Reply-To: [email protected]
>
>New to the list.  I searched through the archives and didn't find this
>question, but if its there and anyone can point me to the thread I'd be
>obliged.  Anyway, here we go:
>
>Performing some relatively deep RIE of SiO2 using an aluminum mask.
>Problem is that mask is eroding more than we'd like (it's not
>sacrificial), even after trying to tweak the recipe.  What we are
>looking at now are ways to 'harden' this aluminum mask such that the
>unwanted milling will be reduced.  The mask is only 15m thick (not
>negotiable), and we need the majority of the mask to remain aluminum for
>functionality after processing.  What we had in mind was enhancing the
>oxidation of the aluminum, making for a thicker Al2O3 layer at the top
>of the metal than the normal native oxide.  Even though it is likely to
>still be milled away, the thicker oxide should last longer and result in
>less overall aluminum milling.  Not sure on thickness requirements, as I
>don't quite know the etch selectivity between Al2O3 and Al, but the
>normal process is currently milling away about 30-40% of the thickness,
>so we'd probably want to find the optimum somewhere between the native
>2.5nm and 250nm.
>
>Here's the problem:  I've been hunting through the literature trying to
>find something for Aluminum like the SiO2 oxidation curves that I could
>find in 5 minutes if I needed to.  Basically, I need to know aluminum
>oxide thickness as a function of time _at elevated temperatures_, and so
>far I've been coming up short.  Would anyone out there be able to point
>me to the appropriate reference, or at least in the right direction?
>
>Also, I've seen that plasma oxidation could be another option instead of
>thermal oxidation, and possibly a better one since it wouldn't require
>temperatures as high, and we could keep the wafer in the closed system
>for both processes.  Any advice on this would be greatly appreciated as
>well.
>
>If providing more info would be helpful, I'd be happy to do so.  Thanks.